
on Computational Economics 
By:  Heinrich, Torsten; Gräbner, Claudius 
Abstract:  Twosided markets are an important aspect of today's economies. Yet, the attention they have received in economic theory is limited, mainly due to methodological constraints of conventional approaches: twosided markets quickly lead to nontrivial dynamics that would require a computational approach, as analytical models quickly become intractable. One approach to this problem is to opt for models that operate on an aggregated level, abstracting from most of the (microlevel) causes of these nontrivial dynamics. Here we revisit a well known equilibrium model by Rochet and Tirole of twosided markets that has taken this approach. Analyzing the model from an agentbased perspective, however, reveals several inconsistencies and implicit assumptions of the original model. This, together with the highly implausible assumptions that are required to make the model analytically tractable, limits its explanatory power significantly and motivates an alternative approach. The agentbased model we propose allows us to study the phenomenon of twosided markets in a more realistic and adequate manner: Not only are we able to compare different decision making rules for the providers, we are also able to study situations with more than two providers.%We find that Thus, our model represents a first step towards a more realistic and policyrelevant study of twosided markets. 
Keywords:  Twosided markets; Network externalities; Agentbased modeling; Simulation; Heuristic decision making; Reinforcement learning; Satisficing; Differential evolution; Evolutionary economics; Market structure; IT economics; Equilibrium dynamics 
JEL:  C61 C62 C63 D4 L14 L15 
Date:  2015–11–13 
URL:  http://d.repec.org/n?u=RePEc:pra:mprapa:67860&r=cmp 
By:  David Sayah (Johannes Gutenberg University Mainz); Stefan Irnich (Johannes Gutenberg University Mainz) 
Abstract:  This paper addresses the discrete pdispersion problem (PDP) which is about selecting p facilities from a given set of candidates in such a way that the minimum distance between selected facilities is maximized. We propose a new compact formulation for this problem. In addition, we discuss two simple enhancements of the new formulation: Simple bounds on the optimal distance can be exploited to reduce the size and to increase the tightness of the model at a relatively low cost of additional computation time. Moreover, the new formulation can be further strengthened by adding valid inequalities. We present a computational study carried out over a set of largescale test instances in order to compare the new formulation against a standard mixedinteger programming model of the PDP, a line search, and a binary search. Our numerical results indicate that the new formulation in combination with the simple bounds is solved to optimality by an outofthebox mixedinteger programming solver in 34 out of 40 instances, while this is neither possible with the standard model nor with the search procedures. For instances in which the line and binary search fail to ?nd a provably optimal solution, we achieve this by adding cuts to our enhanced formulation. 
Keywords:  facility location, dispersion problems, maxmin objective, integer programming 
Date:  2015–11–05 
URL:  http://d.repec.org/n?u=RePEc:jgu:wpaper:1517&r=cmp 
By:  Viktor Tsyrennikov (IMF); Serguei Maliar (Santa Clara University); Lilia Maliar (Stanford University); Cristina Arellano (Federal Reserve Bank of Minneapolis) 
Abstract:  We develop an envelope condition method (ECM) for dynamic programming problems  a tractable alternative to expensive conventional value function iteration. ECM has two novel features: First, to reduce the cost, ECM replaces expensive backward iteration on Bellman equation with relatively cheap forward iteration on an envelope condition. Second, to increase the accuracy of solutions, ECM solves for derivatives of a value function jointly with a value function itself. We complement ECM with other computational techniques that are suitable for highdimensional problems, such as simulationbased grids, monomial integration rules and derivativefree solvers. The resulting valueiterative ECM method can accurately solve models with at least up to 20 state variables and can successfully compete in accuracy and speed with stateoftheart Euler equation methods. We also use ECM to solve a challenging default risk model with a kink in value and policy functions, and we find it to be fast, accurate and reliable. 
Date:  2015 
URL:  http://d.repec.org/n?u=RePEc:red:sed015:1239&r=cmp 
By:  David E. Allen (The University of Sydney, The University of South Australia, Australia); Michael McAleer (National Tsing Hua University, Taiwan; Erasmus University Rotterdam, the Netherlands; Complutense University of Madrid, Spain); Shelton Peiris (The University of Sydney, Australia); Abhay K. Singh (Edith Cowan University, Australia) 
Abstract:  This paper features an analysis of major currency exchange rate movements in relation to the US dollar, as constituted in US dollar terms. Euro, British pound, Chinese yuan, and Japanese yen are modelled using a variety of nonlinear models, including smooth transition regression models, logistic smooth transition regressions models, threshold autoregressive models, nonlinear autoregressive models, and additive nonlinear autoregressive models, plus Neural Network models.The results suggest that there is no dominating class of time series models, and the different currency pairs relationships with the US dollar are captured best by neural net regression models, over the ten year sample of daily exchange rate returns data, from August 2005 to August 2015. 
Keywords:  Non linear models; time series; nonparametric; smoothtransition regression models; neural networks; GMDH shell 
JEL:  C45 C53 F3 G15 
Date:  2015–11–06 
URL:  http://d.repec.org/n?u=RePEc:tin:wpaper:20150125&r=cmp 